Global properties of the mapping between local amino acid sequence and local structure in proteins.

Abstract

Local protein structure prediction efforts have consistently failed to exceed -70% accuracy. We characterize the degeneracy of the mapping from local sequence to local structure responsible for this failure by investigating the extent to which similar sequence segments found in different proteins adopt similar three-dimensional structures. Sequence segments 3-15 residues in length from 154 different protein families are partitioned into neighborhoods containing segments with similar sequences using cluster analysis. The consistency of the sequence-to-structure mapping is assessed by comparing the local structures adopted by sequence segments in the same neighborhood in proteins of known structure. In the 154 families, 45% and 28% of the positions occur in neighborhoods in which one and two local structures predominate, respectively. The sequence patterns that characterize the neighborhoods in the first class probably include virtually all of the short sequence motifs in proteins that consistently occur in a particular local structure. These patterns, many of which occur in transitions between secondary structural elements, are an interesting combination of previously studied and novel motifs. The identification of sequence patterns that consistently occur in one or a small number of local structures in proteins should contribute to the prediction of protein structure from sequence.Most studies of local sequence-structure relationships have involved the tabulation of statistics on sequences that occur in structural motifs of interest (1-4) (Fig. 1A). Our approach (Fig. 1B) is essentially the inverse. Instead of investigating the sequence patterns found in predefined local structural environments, we first identify recurring sequence patterns and then investigate their structural correlates.It is well established that the local sequence-to-structure mapping is not one to one over all of sequence space: identical pentapeptide sequences exist in completely different tertiary structures in proteins (5). Furthermore, local structure prediction efforts consistently fail to exceed -70% accuracy (6), suggesting that the mapping from local sequence to structure is likely to be degenerate for a significant fraction of sequence space. In this paper we characterize the degeneracy of the mapping by determining the number of sequence segments in neighborhoods (regions of sequence space) in which the sequence-to-structure mapping is one to one, one to two, and one to three (Fig. 1B). In previous work (9), we utilized a measure of the distance between sequence profiles generated from multiple sequence alignments to identify sequence patterns that transcend protein family boundaries. A similar distance measure is used in this paper, and the term "segment" below refers to a segment of a profile generated from a multiple sequence alignment.The earlier work focused on the identification and characterization of recurring sequence patterns; the focus of the current paper is on the structural correlates of these patterns. Metho...

Modeling of the D1/D2 proteins and cofactors of the photosystem II reaction center: Implications for herbicide and bicarbonate binding

“…The loop regions are presumably all solvent exposed; it is thus justifiable to use structural fragments from other proteins, even soluble proteins, because loop regions are well characterized by sequences (Han & Baker, 1996). This was done by searching for highly homologous sequences in proteins whose high-resolution structures are available.…”

HMMSTR: a hidden Markov model for local sequence-structure correlations in proteins 1 1Edited by J. Thornton

“…The I-sites (invariant or initiation sites) library consists of an extensive set of short sequence motifs, lengths 3 to 19 motifs obtained by exhaustive clustering of sequence segments from a nonredundant database of known structures (Han & Baker, 1996;Bystroff & Baker, 1998). Each sequence pattern correlates strongly with a recurrent local structural motif in proteins.…”

Sequence context-specific profiles for homology searching

“…By mixing mutation probabilities from the substitution matrices, weighted by posterior probabilities for the corresponding HMM states, HMMSUM achieved considerable improvements in alignment quality when compared with standard substitution matrices (27). We expect such sequence contexts to predict mutation probabilities better than structural environments, because very different sequences with very specific amino acid preferences can adopt similar local structures (28). When all of these sequences are pooled into the same structural environment, the specific amino acid preferences are lost.…”